We followed the diel and summer variations of bacterioplankton production (estimated from [3H] thymidine incorporation) and abundance (direct count with 4′,6-diamidino-2-phenylindole staining) at four depths in a Canadian Shield humic lake. We found production to be highest and most variable in the epilimnion, but the differences between production estimates made at different times were statistically significant in only 5 out of 16 cases. Production differed significantly among depths in 14 out of 18 trials. The renewal time of the bacterial community varied between 1 and 20 days. Bacterioplankton production and primary production were uncorrelated (r = 0.225; p > 0.2), but bacterial production and chlorophyll a concentration were positively correlated (r = 0.816; p < 0.005). Small coccis (~0.3 – ~0.8 μm) represented between 52 and 80% of the bacterial cells in all the samples. There was no correspondence between bacterial production and frequency of dividing cells. Bacterial production and abundance varied inversely throughout the season, bacteria being two to three times less concentrated in the spring, whereas production was two to three times higher. We found no correlation between bacterioplankton production and abundance (r = 0.012; p > 0.5). Our results demonstrate the importance of small-scale sampling and the difficulty with which bacterioplankton production, its spatiotemporal variations, and the relationships between bacteria and phytoplankton can be predicted. Key words: bacterioplankton, abundance, production, spatiotemporal variations, bacterioplankton–phytoplankton relationships.